cancel
Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Discover peer insights or crowdsource your one of a kind Designer question.

Search for and remove people's names from data

Asteroid

Hi,

 

Does Alteryx provide a tool for searching for and flagging up peoples names from a dataset?  We need to be able to search for and remove peoples names that are sent to us in error...

 

Any help/advice is appreciated

 

Many thanks

Highlighted
Magnetar
Magnetar

Hi @fiorano

Yes, they have a couple tools that might suffice:

  • Find/Replace tool: could find a name from a a list of names, and either add a flag or simply remove it
  • Fuzzy Match tool: for more sophisticated name matching

Hope that helps!

John

Asteroid

Hi @JohnJPS

 

Thanks for your prompt reply.....   Thing is we don't have a list of names to compare against...   Im more looking to search through a field and Alteryx to say that "This looks like a persons name" ... is this possible??

Magnetar
Magnetar

Yes, the Fuzzy Match can do that (along with addresses, business names, etc). It's a very sophisticated tool - lots of powerful stuff going on behind the scenes with that one.

Asteroid

hI @JohnJPS

 

I was under the impression you needed a list of names to match against for Fuzzy Match to work....?

Magnetar
Magnetar

The attached workflow uses the .csv from this public Baby Names dataset.  The workflow is based on the Sample available right within Alteryx: I just swapped in the baby names dataset and tweaked accordingly; (e.g. match name instead of company name).  The idea here is it looks at your own dataset and finds things that don't match well with other items in your own data.  You could, for instance, use this idea to curate your own list of approved names (from your own dataset, not the baby names, though that might actually work), and then just deal with the outliers as they arise.

 

Asteroid

HI @JohnJPS

 

Thanks again for this example.  To be honest I'm a touch confused as to how this would work for my needs.  

 

For example, if I had the following data:

 

TextField

Order for John Williams - 89989809824

500 sheets aluminium (repackaged)

Personalised delivery Sandra Roberts - sent 01/08/2017

pallets - raw paperfeed

laser cartridges CYMK (10x30)

 

 

What would be the easiest way to flag the records containing John WIlliams and Sandra Roberts?

 

Many thanks again for your help.

 

 

 

 

 

 

 

 

 

 

Magnetar
Magnetar
I see. I was assuming parsing through a list of names to look for non-names, vs. looking through a list of anything to pick out names.

For your example, I would first try to curate a list of first and/or last names, then focus on using the list to either find/replace or fuzzy search.